Spatial contrast adaptation characteristics of neurones recorded in the cat's visual cortex

Abstract

Spatial contrast adaptation, produced by prolonged exposure to high contrast grating patterns, has become an important psychophysical method for isolating spatial and orientation selective channels in the human visual system. It has been reasonably argued that this adaptation may be fundamentally dependent upon the activity of neurones in the striate cortex. To test the validity of this hypothesis, and several others, we measured the general adaptation characteristics of 144 striate neurones using a stimulus protocol comparable to the typical psychophysical methods. In general, during prolonged high contrast stimulation, the responses of most cells exponentially decayed from a transient peak response to a sustained plateau response; following adaptation, the responses to lower contrasts were depressed relative to the unadapted state but then gradually recovered from the transient depression to a sustained plateau. Such adaptation was a property common to both simple and complex cells (the distributions of the quantitative of adaptation were overlapping); there were however small but reliable differences. We compared the neurophysiological contrast adaptation with two psychophysical estimates of human contrast adaptation (threshold contrast elevation and apparent contrast reduction) and found that the time courses and the magnitudes were quite similar. The effect of contrast adaptation on the spatial frequency tuning was assessed by measuring the contrast response function at several different test spatial frequencies before and after adaptation at the optimum centre frequency. We found that the effect of adaptation decreased as the difference between test and adaptation frequency increased. Grating contrast adaptation has been alternatively described as 'constructive gain control' on the one hand and as 'deleterious fatigue' on the other. We tested the effect of contrast adaptation on the contrast response function and found (a) that adaptation shifts the curves vertically downward parallel to the response axis (thus reflecting a decrease in the maximum rate of firing and a deleterious compression of the response range) and (b) that adaptation shifts the curves horizontally to the right parallel to the contrast axis (thus reflecting a true sensitivity shift of the remaining response range for constructive maintenance of high differential sensitivity around the prevailing background level).(ABSTRACT TRUNCATED AT 400 WORDS)